Well I was surfing Tiktok this morning and ran across someone playing with ChatGPT and wondered what talking to a AI app on your phone would be like. I downloaded and fired up ChatGPT. Grabbed my earbuds and had a chat with AI. I was playing around trying different questions. Well look for yourself.

 

 

Then I continued my chat about economy tuning which got interesting ask about thermodynamics and efficiency of the Cummins engine which gave me some tuning clues I'm going to try this winter. See how far I can continue to push out the MPG numbers.

 

 

Here is another question I asked that agrees with what I already knew but again gotta educate the masses.

 

And this is AI in its infancy. Will go from cool to scary in the blink of an eye. 

Right now the ChatGPT 3.5 I've got is trained with much less information and does create some errors in math and math formulas I found out. The funny part is Elon Musk was part of the start-up of ChatGPT but no longer has any control over the software. 

 

Quote

Does Elon Musk Own ChatGPT?

 

While Elon Musk co-founded ChatGPT and helped to financially back its mother company, OpenAI, in its early years, the ex-Twitter CEO doesn't currently own stakes in the company. Musk officially left OpenAI's Board of Directors in 2018. In a Tweet the tech entrepreneur fired off in 2019, he cited a conflict of interest and prioritization of his other businesses, SpaceX and Tesla, as his primary reasons for distancing himself from the research company.

 

Then who owns ChatGPT...

 

Quote

Who Owns ChatGPT?

 

ChatGPT is owned by OpenAI, an AI research laboratory that was founded in 2015 by Sam Altman, Elon Musk, and other prominent figures including Peter Theil, Ilya Sutskever, Jessica Livingston, Reid Hoffman, Greg Brockman, Wojciech Zaremba, and John Schulman.

 

Differences between 3.5 and 4.0 ChatGPT

 

Quote

Difference 4: ChatGPT 4 is more Accurate, less prone to “hallucinations.”

 

ChatGPT 3.5 was based on the GPT 3.5 engine, which received training on over 175 billion training parameters. ChatGPT 4 took this training one step further, and is purportedly trained on over a trillion parameters. What does this mean? This means that ChatGPT4 will consider those trillion parameters before formulating a response. When ChatGPT 3.5 was initially released, there were instances where people were asking the chatbot ways to hack into banking websites and known vulnerabilities in other popular sites. This will not be possible in ChatGPT4.